Private Redaction

Redact sensitive info from images, screenshots and text, privately in your browser, with a verifiable private-AI assist.

Private Redaction black-outs sensitive information in images, screenshots and text, right in your browser. Nothing is uploaded to redact it: your image or text is opened and the redacted result is produced on your own device. Three ways to use it: - Right-click an image and choose "Redact this image". - Click the toolbar icon and drag to capture any region of a page (for example, screenshot a balance but cover the numbers). - Right-click selected text and choose "Redact selected text". In the workspace you can always drag on the image to black out anything, and click a box to remove it. Auto-detect adds redactions over things it finds using built-in pattern rules (emails, phone / card / account numbers, IBANs, and more) that run entirely on your device. Review every detection in a checklist and untick anything you want to keep. Optional private-AI assist: For fuzzy categories like names, addresses and organisations, Private Redaction can call a private AI that runs inside a sealed Nillion enclave (AMD SEV-SNP secure hardware). Only the extracted text is sent, straight from your browser to the enclave, using a short-lived token minted from a public key. Our own servers never receive your text on that path, and you get a verifiable hardware-attestation receipt for each run. Private by design: - Redaction and OCR run on your device. - Manual redaction and the local rules need no network at all. - The AI step sends only extracted text, direct to the sealed enclave, and it isn't stored. - The output is a flattened image: the covered pixels are gone, not hidden. Automated redaction can miss things or cover too much, so always review the result before you rely on it or share it. Private Redaction is a free, open-source, experimental tool. Learn more and read the source at privateredact.app.